Commit history

1. Commits on Apr 22, 2026

   • Performance Overhaul: Tuned entropy routing, lowered Surya DPI to 144, and eliminated redundant table extraction for a 7-10% speed gain.

     alemoriv

     committed
     4f11fe2View commit details

     Copy full SHA for 4f11fe2

     Browse repository at this point

   • Integrated Surya OCR math extraction with surgical para preservation, K-Means sorting, and table optimizations.

     alemoriv

     committed
     8110f2cView commit details

     Copy full SHA for 8110f2c

     Browse repository at this point

   • Refactor pdf-ocr.py into a modular parallelized pipeline with graph-based layout

     H4Z33

     committed
     467f574View commit details

     Copy full SHA for 467f574

     Browse repository at this point

   • fixed version ocr-surya stable

     H4Z33

     committed
     0fdee39View commit details

     Copy full SHA for 0fdee39

     Browse repository at this point
2. Commits on Apr 21, 2026

   • Enhance math extraction with geometric shard merging and structural anti-outlier filtering; cleanup OCR artifacts

     alemoriv

     committed
     8d84dd8View commit details

     Copy full SHA for 8d84dd8

     Browse repository at this point
3. Commits on Apr 16, 2026

   • feat: integrate Surya-OCR for high-fidelity LaTeX math extraction and programmatic inline notation recovery

     alemoriv

     committed
     ae9498eView commit details

     Copy full SHA for ae9498e

     Browse repository at this point

   • feat: implement localized inline math surgical replacement for OCR pipeline

     alemoriv

     committed
     01e2a23View commit details

     Copy full SHA for 01e2a23

     Browse repository at this point

   • .

     alemoriv

     committed
     77773cbView commit details

     Copy full SHA for 77773cb

     Browse repository at this point

   • first commit

     alemoriv

     committed
     14bf532View commit details

     Copy full SHA for 14bf532

     Browse repository at this point